Upload README.md with huggingface_hub

README.md

@@ -22,57 +22,6 @@ language:
 
 [![Qwen Chat](https://img.shields.io/badge/%F0%9F%92%9C%EF%B8%8F%20Qwen%20Chat%20-536af5)](https://chat.qwen.ai)
 
-> [!Note]
-> This repository contains model weights and configuration files for the post-trained model in GGUF format.
->
-> These artifacts are intended for llama.cpp-style runtimes and other GGUF-compatible inference stacks.
-
-Following the February release of the Qwen3.5 series, we're pleased to share a `TQ3_4S` GGUF release of Qwen3.6-27B. Built on the upstream Qwen3.6-27B model and converted through the `unsloth/Qwen3.6-27B-GGUF` path, this release is aimed at strong local inference efficiency while preserving the stability and real-world coding utility of the base model.
-
-## Qwen3.6 Highlights
-
-This release delivers substantial upgrades, particularly in
-
-- **Agentic Coding:** the model now handles frontend workflows and repository-level reasoning with greater fluency and precision.
-- **Thinking Preservation:** Qwen introduced the option to retain reasoning context from historical messages, reducing overhead during iterative work.
-
-![Benchmark Results](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3.6/Figures/qwen3.6_27b_score.png)
-
-For the full upstream write-up, see the Qwen blog post: [Qwen3.6-27B](https://qwen.ai/blog?id=qwen3.6-27b).
-
-## Model Overview
-
-- Type: Causal Language Model with Vision Encoder
-- Training Stage: Pre-training and Post-training
-- Architecture: `qwen35`
-- Parameters: `27B`
-- Layers: `64`
-- Embedding dimension: `5120`
-- FFN dimension: `17408`
-- Hidden layout: `16 × (3 × (Gated DeltaNet -> FFN) -> 1 × (Gated Attention -> FFN))`
-- Gated DeltaNet heads: `48` for `V`, `16` for `QK`, head dim `128`
-- Gated Attention heads: `24` for `Q`, `4` for `KV`, head dim `256`
-- RoPE dim: `64`
-- Native context: `262,144`
-
-## Benchmark Results
-
-For the full upstream benchmark tables, refer to the official Qwen model card:
-
-- [Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B)
-
-Selected upstream headline results for the base model:
-
-- `SWE-bench Verified`: `77.2`
-- `Terminal-Bench 2.0`: `59.3`
-- `SkillsBench Avg5`: `48.2`
-- `GPQA Diamond`: `87.8`
-- `AIME26`: `94.1`
-- `MMMU`: `82.9`
-- `AndroidWorld`: `70.3`
-
-These are upstream base-model results, not local GGUF quant results.
-
 ## TQ3_4S Release
 
 This repository packages the model as a TurboQuant `TQ3_4S` GGUF for local deployment.
@@ -103,6 +52,12 @@ Prompt processing:
 - `PPL = 6.2452 +/- 0.16138`
 - `prompt eval = 712.02 tok/s`
 
+16 GB VRAM fit checks on RTX 5060 Ti with the recommended KV settings:
+
+- `32k` context fits
+- `64k` context fits
+- `128k` context does not fit
+
 ## Runtime Notes
 
 - Use a TurboQuant-capable llama.cpp build for best performance.
@@ -112,6 +67,31 @@ Prompt processing:
 - Upstream guidance recommends keeping at least `128K` context when possible for reasoning-heavy workloads. On smaller local GPUs, reduce context as needed to fit memory.
 - Upstream default sampling guidance differs between thinking and non-thinking mode; follow the official Qwen card if you are trying to reproduce base-model behavior.
 
+## Recommended llama.cpp Settings
+
+Default prompt-processing settings on 16 GB:
+
+```bash
+llama-bench \
+  -m Qwen3.6-27B-TQ3_4S.gguf \
+  -ngl 99 \
+  -ctk q4_0 \
+  -ctv tq3_0 \
+  -fa 1 \
+  -p 2048 -n 0 -r 3
+```
+
+Default chat/server settings:
+
+```bash
+llama-server \
+  -m Qwen3.6-27B-TQ3_4S.gguf \
+  --host 127.0.0.1 --port 8080 \
+  -ngl 99 -c 4096 -np 1 \
+  -ctk q4_0 -ctv tq3_0 -fa on \
+  --jinja
+```
+
 ## Example
 
 ```bash
@@ -128,6 +108,47 @@ Build/runtime:
 git clone https://github.com/turbo-tan/llama.cpp-tq3
 ```
 
+## Qwen3.6 Base Model
+
+> [!Note]
+> The upstream Qwen repository contains model weights and configuration files for the post-trained model in the Hugging Face Transformers format.
+>
+> Those upstream artifacts are compatible with Hugging Face Transformers, vLLM, SGLang, KTransformers, and related runtimes.
+
+Following the February release of the Qwen3.5 series, Qwen describes Qwen3.6 as the first open-weight Qwen3.6 variant, built for stronger stability and real-world utility.
+
+### Qwen3.6 Highlights
+
+- **Agentic Coding:** the model handles frontend workflows and repository-level reasoning with greater fluency and precision.
+- **Thinking Preservation:** the model family retains reasoning context across historical turns to reduce overhead during iterative work.
+
+![Benchmark Results](https://qianwen-res.oss-cn-beijing.aliyuncs.com/Qwen3.6/Figures/qwen3.6_27b_score.png)
+
+### Model Overview
+
+- Type: Causal Language Model with Vision Encoder
+- Training Stage: Pre-training and Post-training
+- Architecture: `qwen35`
+- Parameters: `27B`
+- Layers: `64`
+- Embedding dimension: `5120`
+- FFN dimension: `17408`
+- Hidden layout: `16 × (3 × (Gated DeltaNet -> FFN) -> 1 × (Gated Attention -> FFN))`
+- Gated DeltaNet heads: `48` for `V`, `16` for `QK`, head dim `128`
+- Gated Attention heads: `24` for `Q`, `4` for `KV`, head dim `256`
+- RoPE dim: `64`
+- Native context: `262,144`
+
+### Selected Upstream Benchmark Highlights
+
+- `SWE-bench Verified`: `77.2`
+- `Terminal-Bench 2.0`: `59.3`
+- `SkillsBench Avg5`: `48.2`
+- `GPQA Diamond`: `87.8`
+- `AIME26`: `94.1`
+- `MMMU`: `82.9`
+- `AndroidWorld`: `70.3`
+
 ## Sources
 
 - Upstream base model: [Qwen/Qwen3.6-27B](https://huggingface.co/Qwen/Qwen3.6-27B)